DESCRIPTION: (Applicant's Abstract) Oxford House (OH), founded in 1975, illustrates a community-based approach toward substance abuse abstinence. Unlike traditional hospital care and therapeutic communities, which involve the use of professionals and have limitations on length of stay, Oxford House offers a community where residents can live without the involvement of professional treatment staff and where there are no time restrictions on residency (Oxford House Manual, 1988). Because there is no maximum stay nor involvement of professionals, OH may offer a cost effective alternative to more traditional approaches to substance abuse recovery whereby residents may have a greater opportunity to develop necessary skills and increase their self-efficacy toward maintaining abstinence. An Oxford House communal living experience offers residents abstinence social support networks. To the extent that recovering substance abusers invest or commit themselves to these networks (Longabaugh et al., 1993), it would be expected that support for abstinence from similar others living with an Oxford House resident would strengthen that person's self-efficacy toward substance abuse abstinence. Thus, theoretically, abstinence support may strengthen one's own self-efficacy for abstinence and as social investment in abstinence support becomes stronger, increases in the person's abstinence self-efficacy may promote substance abuse abstinence. The proposed study will use an accelerated longitudinal design to examine the relation between abstinence support (moderated by social investment), development of self-efficacy, and successful abstention from substance use in a national sample of Oxford House residents. In this study, we will recruit a sample of 151 Houses, each with 8 residents on average, yielding approximately 1,208 residents. The participants will be interviewed at the initial baseline phase and tracked for one year. Follow-up interviews will be conducted at 3-month intervals for a total of 5 assessments per individual. The specific aims to evaluate new dimensions of Longabaugh et al.'s (1993) model are: 1) to examine whether increases in self-efficacy predict successful substance use outcomes (i.e., abstinence status, reduction in symptoms, and quantity of use); and, 2) to examine whether changes in the degree of abstinence support received, moderated by social investment, promote these increases in self-efficacy.